Cutaneous painting with reactive haptens induces contact sensitivity (CS) responses that are in vivo examples of T cell immunity. In contrast, high dose i.v. administration of the hapten can induce tolerance. We investigated the effect of IL-12 on reversal of this tolerance and attempted to determine in vitro the mechanism of this reversing effect by measuring proliferation and IFN-γ production by CS effector T cells stimulated with hapten-conjugated APC, and we also measured CS ear swelling in vivo. The in vitro responses of T cells to hapten-APC became absent in tolerized mice, paralleling impaired in vivo CS responses. Addition of IL-12 to cultures manifesting this fully established in vitro tolerance completely restored impaired responses of tolerized T cells. The reversing effects of IL-12 were not blocked by anti-IFN-γ mAb, but were blocked by mAbs against B7-1, more strongly by anti-B7-2, and by both Abs together. Additional in vivo ear-swelling response experiments confirmed the reversing effects of IL-12 on established tolerance. To examine whether the IL-12 effect depended on stimulation of IFN-γ, we directly injected IFN-γ into tolerized mice. This partially mimicked but did not fully reconstitute the effects of IL-12. In summary, IL-12 abrogation of established tolerance of CS may have been partially due to endogenous production of IFN-γ, but appeared mainly due to direct activation of the tolerized T cells by affecting signaling through costimulatory molecules B7-1 and B7-2.
Contact sensitivity (CS)3 is induced by skin painting with reactive hapten and is a classical manifestation of in vivo T cell-mediated immunity. In contrast, i.v. administration of a high dose of similarly reactive, but water-soluble, hapten induces tolerance, i.e., Ag-specific unresponsiveness for CS (1). This i.v. tolerogenic treatment induces down-regulatory T cells that impair responses of CS effector cells when adoptively cotransferred to naive recipients in cell-mixing experiments with the effector cells (2, 3, 4). In the current study, we tested the effect of IL-12, a critical positive-acting cytokine for Th1 responses, on reversing CS tolerance induced by prior i.v. administration of hapten. We demonstrate the complete reversal of this established high Ag dose CS tolerance by IL-12 treatment, and that reversal of tolerance by IL-12 required both B7-1 and B7-2 costimulatory molecules. We employed correlating in vitro and in vivo assays of Ag-specific Th1 T cell function to achieve these findings.
IL-12 is a heterodimeric Th1-associated cytokine (5, 6) that consists of disulfide-linked 35- and 40-kDa chains (7) and is produced by macrophages, B cells, and possibly other cells (8, 9). IL-12 regulates the growth and function of T cells (10, 11) and especially regulates the development of Th1 cells by stimulating the production of IFN-γ (6, 12, 13). Prior results showed that IL-12 promoted resistance to Leishmania (14, 15) and other parasites (16, 17) and inhibited airway hypersensitivity in an asthma model (18) by directing immunity toward Th1 and away from Th2 responses, and in other systems similarly promoted cell-mediated immunity against bacteria, intracellular parasites, and tumors (19, 20, 21).
An important role of IL-12 also has been reported in induction and elicitation of CS (22, 23). In particular, neutralization of IL-12 by in vivo administration of mAb impaired generation of CS responses, switching the development of CS immunity to induction of tolerance (22). These results suggested that IL-12 also was a positive regulator of the development of CS responses. Furthermore, in vivo injection of IL-12 at the induction reversed UV-induced tolerance (23, 24). Tolerance induced by UV irradiation was reported to be due to IL-10 released from UV-irradiated local keratinocytes (25, 26). Accordingly, anti-IL-10 mAb treatment prevented UV tolerance (25, 27). Furthermore, tolerance of CS was induced by IL-10 administration at induction, suggesting a role for Th2 cytokines in CS tolerance (28). The above studies implied that a cytokine imbalance between Th1 vs Th2 cells was responsible for development of UV-induced tolerance in CS, and IL-12 could prevent the induction of tolerance by antagonizing Th2 cytokines. However, compared with the induction of tolerance, there is little evidence that these inhibitory Th2 cytokines also are involved in the maintenance of established tolerance. Accordingly, most of the effects of IL-12 in CS have been investigated at the time of immunization by injecting or neutralizing IL-12 during primary immune responses (i.e., the induction phase), but not at the stimulation of secondary immune responses (i.e., elicitation phase) of CS. Thus, the effects of IL-12 on this established tolerance that was induced by i.v. administration of high doses of reactive hapten, and the molecular mechanisms that might underlie the reversal of established tolerance by IL-12 were still unclear.
In the current study, we analyzed the underlying molecular mechanisms in previously induced and established tolerance of CS, including the effect of IL-12 on a form of established CS tolerance not studied previously. Fully established CS tolerance was induced by i.v. administration of high doses of reactive hapten, and then CS immunization was attempted. Also, we employed newly developed in vitro assays of T cell proliferation and IFN-γ production by CS effector T cells responding to hapten-conjugated APC that paralleled in vivo tolerance of CS ear-swelling responses. We found that in vitro addition of IL-12 largely reversed established tolerance of in vitro responses, and that these effects of IL-12 required both B7-1 and B7-2 molecules, since Abs to both B7-1 and B7-2 blocked the ability of IL-12 to reverse CS tolerance. Furthermore, administration of IL-12 in vivo also reversed this established tolerance of CS. In contrast, IFN-γ administration in vivo only partially mimicked the effects of IL-12.
In summary, these results suggest that IL-12 may have reversed established CS tolerance by direct activation of tolerized CS effector T cells in which both B7-1 and B7-2 molecules are required for production of IFN-γ and proliferation. Thus, one of the underlying mechanisms that might be involved in the reversal of tolerance by IL-12 might be a synergistic action of this cytokine with B7-1 and B7-2 molecule-mediated signals in CS effector T cells.
Materials and Methods
Male 6- to 8-wk-old CBA/J mice were obtained from The Jackson Laboratory (Bar Harbor, ME), kept in filter-topped microisolators, and rested 1 to 2 wk before use.
Picryl chloride (PCl) (trinitrophenyl (TNP) chloride) was obtained from Chemica Alta (Edmonton, Alberta, Canada), recrystallized from methanol/H2O, and stored in a light-protected desiccator at room temperature. Oxazolone (4-ethoxymethylene-2-phenyloxazolone (OX)) and trinitro-benzene sulfonic acid (TNBSA) were purchased from Sigma Chemical Co. (St. Louis, MO) and Eastman Kodak Co. (Rochester, NY), respectively.
Abs and cytokines
Hybridoma cells producing anti-mouse IL-10 mAb (clone; JES5–2A5.1.1, rat IgG1) were obtained from the American Type Culture Collection (Rockville, MD) with the permission of DNAX Research Institute (Palo Alto, CA) and were maintained in RPMI 1640 containing 10% heat-inactivated FCS (Gemini, Calabasas, CA), 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, and 25 mM HEPES. Purified mAb to IL-10 was prepared by passage of hybridoma culture supernatants over a protein G-Sepharose column (Pharmacia, Piscataway, NJ). mAbs to TGF-β2 (clone; 1D11.16, mouse IgG1) (29) was provided by Dr. Wendy Waegel at Celtrix Laboratories (Santa Clara, CA). Neutralizing mAb to murine B7-1 (clone; 1G10, rat IgG2a) and B7-2 (clone; 2D10, rat IgG2b) were provided by Dr. Denise Faherty at Hoffmann-La Roche (Nutley, NJ). Murine rIL-12 and sheep polyclonal Ab to murine IL-12 were generous gifts from Dr. Stanley F. Wolf at Genetics Institute (Cambridge, MA). Neutralizing Abs to murine IFN-γ (clone; XMG1.2, rat IgG1) and murine IL-2 (clone; C300–1A12) were provided by Dr. Robert Coffman at DNAX Research Institute. Murine rIFN-γ (4.6 × 106 U/mg) was provided by Toray Industry, Inc. (Kanagawa, Japan).
Hapten conjugation of splenic APC
Normal mouse spleen cells were treated with 100 μg/ml of mitomycin C (Sigma Chemical Co.) at 37°C for 30 min. Then, after washing with PBS, the cells were incubated with 10 mM TNBSA (pH 7.2) or 6 mM Oxazolone in HBSS (pH 7.2), at 37°C for 10 min. Oxazolone was first dissolved in 100 μl ethanol, heated to 37°C, diluted with warmed HBSS, and used immediately.
In vitro measurement of CS immune lymph node cell (LNC) proliferation
A single cell suspension of LNC (axial, brachial, and inguinal) was prepared under aseptic conditions, and 4 × 105 LNC were incubated in flat-bottom 96-well microplates (Falcon, Oxnard, CA) with various numbers of hapten-conjugated syngeneic normal spleen cells as APC in 0.2 ml of RPMI 1640 containing 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, 25 mM HEPES, 5 × 10−5 M 2-ME, and 10% FCS for 72 h. [3H]thymidine (1 μCi/well; Amersham, Aylesbury, U.K.) was added for the last 6 of 72 h of culture incubation, and [3H]thymidine incorporation was determined by beta liquid scintillation counting. Ssupernatant was also collected at 48 h for assay of IFN-γ production and was stored at −20°C until use.
ELISA detection of IFN-γ in culture supernatants
Quantitative ELISA of IFN-γ employed two different mAb specific for mouse IFN-γ, according to the manufacturer (PharMingen, San Diego, CA). Briefly, wells of 96-well microtiter plates (Corning, Corning, NY) were coated with 1 μg/ml of capture mAb (clone R4-6A2) in 0.1 M NaHCO3 (pH 8.3) at 4°C overnight. Following blocking with 3% dry milk in PBS at room temperature for 2 h, samples and standard recombinant mouse IFN-γ dilutions (Genzyme, Cambridge, MA), were added and incubated overnight at 4°C. Then, 0.5 μg/ml of biotinylated detection mAb to mouse IFN-γ (clone XMG1.2) and subsequently 1/3000 diluted horseradish peroxidase-conjugated streptavidin (Vector Laboratories, Burlingame, CA) were added to probe for IFN-γ, and then TMB (tetramethylbenzidine), the peroxidase substrate, and TMB one-component stop solution (Kirkegaard and Perry, Gaithersburg, MD) were used to obtain ODs determined at 450 nm.
Contact sensitization and i.v. tolerance induction
Positive control mice were contact sensitized by topical skin application of 5% PCl (TNP-chloride) or 3% OX (150 μl) in an ethanol-acetone mixture (3:1) to the shaved abdomen and four footpads. For prior TNBSA tolerization, separate mice initially received i.v. 0.35 ml of 1% TNBSA in PBS (pH adjusted to 7.2 with 1 M NaOH) on days 0 and 3. Then, to evaluate tolerance, these mice were contact sensitized by skin painting on day 7 with the same or a different hapten (TNP vs OX), used for tolerance. Since no water-soluble salt of OX existed for OX tolerization similar to TNBSA, OX hapten-conjugated syngeneic spleen cells were prepared (1). Briefly, 1 mg of OX initially dissolved in a small volume of ethanol was brought to 1 ml with HBSS (pH 7.2), and spleen cells were incubated at 25°C for 1 h in this solution at 107 cells/ml. After washing with PBS, 5 × 107 of the OX hapten-conjugated cells/mouse were injected i.v. on day 0. Then, to test tolerance, these OX-injected mice were contact sensitized on day 7 by applying 3% OX or 5% PCl to the shaved abdomen and four footpads. To determine the CS ear-swelling responses of these i.v. tolerized and then CS-immunized animals, the mice were challenged with 10 μl of 0.8% PCl or OX in olive oil onto each side of both ears, and subsequent increases in ear thickness were measured 24 h after challenge, using a dial thickness gauge (Langenmessgerate, Berlin, Germany).
In some experiments, murine rIL-12 (0.5 μg/mouse) or IFN-γ (100 or 500 U/mouse) was injected i.p. into various mice 24 and 1 h before ear challenge.
Data were analyzed using Student’s t test (two-tailed) for independent samples, and p < 0.05 was taken as the level of significance.
In vitro quantitation of Ag-specific tolerance induced in vivo
To study established Ag-specific tolerance of CS in vitro, we examined proliferative responses of LNC from mice that were tolerized by i.v. injection of hapten in vivo and then were tested in vivo for resulting tolerance by attempted CS induction via hapten skin painting. The LNC from mice tolerized i.v. with TNBSA before TNP contact sensitization showed complete in vitro unresponsiveness of T cells to proliferation induced by various doses of TNP-APC compared with LNC of mice that were just contact sensitized with TNP and not previously i.v. tolerized (Fig. 1,A). This T cell unresponsiveness in vitro was Ag specific, since mice tolerized instead with OX, a non-cross-reacting hapten, and then similarly contact sensitized with TNP, in contrast, had intact in vitro proliferative responses to TNP-APC (Fig. 1 A).
In similar, but reciprocal, experiments using this different OX hapten, we obtained similar results of Ag-specific unresponsive proliferation in vitro (Fig. 1,B). Thus, LNC from OX-tolerized and then OX contact-sensitized mice were unresponsive in vitro to OX-APC, but, in contrast, LNC from TNBSA-tolerized and then similarly OX contact-sensitized mice had intact proliferative responses to OX-APC in vitro (Fig. 1 B). Therefore, Ag-specific tolerance induced in vivo by i.v. injection of high doses of hapten could be demonstrated in vitro. These in vitro results of unresponsive T cell proliferation were exactly analogous to previously reported tolerized in vivo ear-swelling responses, in which decreased hapten-specific ear-swelling responses were induced by similar i.v. administration of high doses of hapten before subsequently testing the tolerance by contact sensitization via skin painting with reactive hapten (1, 2, 30).
IL-12 added in vitro reverses the reduced T cell proliferation and IFN-γ production of tolerized mice
Since IL-12 has a strong ability to stimulate Th1 cells (6, 12, 13) that usually mediate CS (31, 32, 33), we determined whether IL-12 could reverse established high dose i.v. hapten Ag tolerance. In vitro T cell proliferative responses to hapten-APC were determined in the presence or the absence of IL-12 added in vitro. Addition of IL-12 in vitro over 3 days caused restoration of proliferative responses to increasing doses of TNP-APC in previously tolerized T cells (Fig. 2). The reversal of tolerance by IL-12 was Ag specific, since these proliferative responses were not induced to the inappropriate hapten OX-APC (Fig. 2). This ability of IL-12 to reverse Ag-specific tolerance of T cell proliferative responses was neutralized completely by addition of polyclonal anti-IL-12 Ab (Fig. 3), indicating that the restoring activity was due to IL-12 itself and not to LPS contamination.
We also noted that in vitro production of IFN-γ by CS effector T cells in response to TNP-APC was decreased substantially in LNC from TNBSA-tolerized mice (Fig. 4, group A vs group B), suggesting that in vitro Ag-specific established tolerance was also observed for T cell cytokine production in addition to T cell-proliferative responses (Figs. 1 and 2). Again, IL-12 largely restored diminished IFN-γ production (Fig. 4, group D). Thus, although addition of IL-12 alone, without Ag stimulation, caused a very small increase in the basal production of IFN-γ (Fig. 4, group C), addition of IL-12 with appropriate TNP-APC produced a strong reversal of T cell IFN-γ production to 52 ng/ml (Fig. 4, group D). This reached 60% of the level of IFN-γ produced by positive CS effector LNC from nontolerized mice (Fig. 4, group A). In contrast, in vitro addition of IL-12 to the i.v. TNBSA cells tolerized in vivo and then stimulation with the inappropriate hapten OX-APC in vitro led to the production of only 7 ng/ml IFN-γ (Fig. 4, group E), similar to the small amounts produced by tolerized LNC without Ag stimulation, even in the presence of IL-12 (Fig. 4, group C). We also demonstrated that the effect of IL-12 on IFN-γ production by tolerized T cells responding to hapten-APC was dose dependent (Fig. 5, groups A–E) and again was neutralized by anti-IL-12 (Fig. 5, group E vs group F). These experiments suggested that IL-12 added in vitro restored Ag-specific hyporesponsive T cell proliferation and IFN-γ production by LNC from mice with established in vivo tolerance induced by high doses of hapten i.v.
IL-12 reversal in vitro of established tolerance of CS is not mediated by IFN-γ and requires costimulatory molecules, B7-1 and B7-2
Next we tried to determine the mechanisms of IL-12 reversal of tolerance of CS effector cells. Since IL-12 was known to have a strong ability to promote IFN-γ production (5, 6, 34), which is an important Th1 cytokine for mediating CS responses (4, 31, 32, 33), we investigated whether the restoring effect of IL-12 was mediated by endogenously produced IFN-γ by experiments in which a neutralizing Ab to IFN-γ was added in vitro. As shown in Figure 6, the restoring effect of IL-12 in proliferative responses was not affected by the presence of anti-IFN-γ mAb and was only partially inhibited to a minor degree by anti-IL-2 mAb.
We then sought to determine whether the effect of IL-12 was mediated via costimulatory molecules such as B7-1 and B7-2. It was reported that IL-12 can activate T cells directly without any production of IFN-γ (34, 35), and that costimulatory molecule-mediated signals had an important role in this direct action of IL-12 on T cells (33, 36). Thus, we added neutralizing mAbs to B7-1 and B7-2 to cultures of the previously tolerized cells in the presence of IL-12 and found that the ability of IL-12 to induce reversal of tolerance and activation of T cells was blocked in vitro by anti-B7-1 and B7-2 mAbs when we measured production of IFN-γ (Fig. 7,A, left) and proliferation (Fig. 7,A, right) of LNC. Interestingly, both B7-1 and B7-2 molecules seemed to be necessary for the restoring effects of IL-12. However, if we looked at hapten-specific proliferative responses and IFN-γ production in normally immune LNC from CS mice that were not tolerized (Fig. 7,B), the B7-1 molecule appeared to contribute little to hapten-specific proliferative responses in these assays (Fig. 7,B, right, group B), but B7-2 was involved in proliferative and IFN-γ responses of these CS immune mice (Fig. 7 B, group C). These results suggested that selective usage of B7-1-mediated signals was involved in the reversal of proliferative responses of tolerized LNC by IL-12, and B7-2-mediated signals also were required for the reversal effects on CS tolerance of IL-12 as well as for hapten-specific activation of normally immune LNC.
IL-12 in vivo reverses hyporesponsive CS ear-swelling responses of tolerant mice
Since IL-12 reversed tolerance in vitro, we investigated whether IL-12 given in vivo also could reverse tolerance of CS by measuring alterations of impaired CS ear-swelling responses. IL-12 was injected twice into the mice, 24 and 1 h before Ag challenge on the ears. Tolerance was already established, having been induced by TNBSA i.v. on days 0 and 3, and then CS immunization was attempted on day 7. IL-12 administration to contact-sensitized mice that had not been tolerized did not affect CS responses (Fig. 8, groups C and D). Also, background ear-swelling responses in nonimmune mice were unaffected (Fig. 8, groups A and B). In contrast, IL-12 caused significant restoration of hyporeactive CS ear-swelling responses in mice with established tolerance (Fig. 8, group E vs group F). These in vivo results confirmed the in vitro results reported above and clearly demonstrated that IL-12 also reversed established tolerance of CS ear-swelling responses in vivo.
In vivo administration of IFN-γ partially reverses hyporeactive CS ear-swelling responses in tolerized mice
The above results showed that IL-12 restoration in vivo of hyporeactive CS ear-swelling responses correlated with in vitro restoration of T cell proliferative responses and IFN-γ production in tolerized mice. Since IL-12 is known to have a strong ability to promote IFN-γ production (5, 6, 34), which is an important Th1 cytokine for mediating CS responses (4, 31, 32, 33), we examined whether the in vivo restorative effect of IL-12 may have been due to its ability to induce production of IFN-γ. Thus, to imitate this effect we directly injected IFN-γ into mice in vivo with the same timing as that used for IL-12 administration, i.e., 24 and 1 h before Ag ear challenge.
We choose this approach rather than administration of anti-IFN-γ mAb together with IL-12 because IFN-γ is known to be a critical cytokine for elicitation of the CS effector responses we studied (4, 31, 32, 33). Thus, we postulated that anti-IFN-γ mAb administration would disturb elicitation of all the fundamental CS effector responses we were measuring, and thus that these effects might not necessarily be mediated by administration of IL-12. Accordingly, IFN-γ administration (100 or 500 U) in vivo had little effect on CS ear-swelling responses of mice that were immunized but were not tolerized, or tended to very slightly decrease CS responses (Fig. 9, group B vs group C or D). In contrast, statistically significant but small and partial restoration of CS ear-swelling responses was observed when the same doses of IFN-γ were administrated to mice with established tolerance and prior Ag ear challenge (Fig. 9, group E vs group F or G). However, restoration of ear-swelling responses by IFN-γ (group F, 42% of positive nontolerized controls; group G, 40% of positive nontolerized controls) was much smaller than restoration by IL-12 (78% of positive nontolerized controls) (Fig. 8, group F). These results suggested that increased endogenous production of IFN-γ may at best have been only partially responsible for the restoration of hyporeactive CS ear-swelling responses observed in tolerized mice injected with IL-12.
Extensive prior studies have attempted to understand the effects of IL-12 on the development of Th1 and Th2 cells (6, 13, 34) and on the respective immunoprotective and deleterious roles of Th1 vs Th2 cells in resistance to parasites (13, 14, 15, 16, 17, 20) or to bacterial infections (12, 19). However, the effects of IL-12 on specific immunologic tolerance have not been elucidated fully. In contrast to most prior studies that examined the induction of immunologic tolerance, in the current study we demonstrate the ability of IL-12 to reverse established T cell tolerance of previously induced cutaneous CS efferent responses, which are complex in vivo effector cellular immune responses that are elicited by Th1 effector cells (4, 37). The i.v. administration of high doses of hapten or of hapten-coupled syngeneic cells before skin painting immunization to induce CS led to hapten Ag-specific unresponsiveness (tolerance) of 24-h in vivo ear-swelling responses to hapten challenge on the ears in an attempt to elicit CS, (1, 2, 30) and also tolerance to T cell proliferation in vitro. In our study, both these manifestations of established tolerance were reversed by administration of IL-12. While manipulations that alter induction of tolerance are of great interest, manipulations that can alter established tolerance are potentially of greater biologic and certainly clinical importance because it is established tolerance that occurs clinically sometimes in cancer and in infections with immunodeficiency, where there obviously is no ability to deal with the induction of tolerance.
Since IL-12 is a strong inducer and promoter of Th1 responses, we investigated whether IL-12 could influence this in vivo high dose hapten-induced established tolerance of Th1 CS responses. Importantly, we tried to understand the molecular mechanisms of this hapten-induced tolerance by developing an in vitro assay of CS tolerance. Using this in vitro system, we determined the state of activation of CS effector T cells by measuring hapten-specific T cell proliferative responses and also IFN-γ production. We showed that CS effector T cells proliferate and produce IFN-γ in response to increasing doses of specific hapten-conjugated APC. Also, we found that T cells from high dose i.v. hapten-treated, tolerized animals showed strongly reduced proliferative responses and reduced in vitro IFN-γ production to specific hapten-APC (Fig. 1). This in vitro unresponsiveness correlated with inhibited CS ear-swelling responses that follow identical tolerogenesis in vivo (1, 2, 30). Importantly, we showed further that IL-12 restored impaired in vitro Ag-specific proliferative responses and IFN-γ production of LNC from mice with high dose hapten tolerance ( Figs. 2–6), and these IL-12 effects required costimulatory molecule-mediated pathways (Fig. 7). Further, we confirmed these in vitro results by showing that in vivo administration of IL-12 restored this same tolerance of CS itself (Fig. 8). Reversal by IL-12 clearly showed that this tolerance was not due to T cell deletion and further suggested that tolerance may not have simply been due to anergy (38).
Since development of tolerance in a different system that was induced by UV irradiation was reported to be mediated by yet another mechanism, namely down-regulatory Th2 cytokines such as IL-10 (25, 26), and recent studies on CS due to CD8 effector T cells (39) also demonstrated down-regulatory CD4 Th2 cells (40), it was possible that the mechanism of IL-12 reversal of established tolerance involved the induction of Th1 cytokines such as IFN-γ to antagonize the well-known down-regulatory effects of Th2 cytokines. However, in contrast to UV-induced tolerance (25, 26) or inhibition of CD8 CS (40), IL-10 was not involved in the induction and maintenance of tolerance that we induced in CD4 CS (37), since in vivo administration of neutralizing mAb against IL-10 during induction of this tolerance did not prevent the development of tolerance (our unpublished observations). In addition, treatment of in vitro cultures of established tolerized cells from our system with neutralizing mAbs against IL-4 and IL-10 did not reverse tolerance (our unpublished observations), indicating that the established high dose i.v. hapten-induced tolerance that we have employed is different from UV-induced tolerance, and once tolerance has been established, Th2 cytokines such as IL-4 and/or IL-10 might not be involved in the maintenance of tolerance. Thus, the mechanisms of reversal of established tolerance by IL-12 that we observed here seemed to be completely different from that of prevention of UV-induced tolerance by IL-12, i.e., direct activation of CS effector cells via costimulatory molecules in our system vs antagonizing effects of IL-12 against suppressive Th2 cytokines.
In many in vivo experiments in which exogenous IL-12 has been used to reverse Th2 responses, it has been shown that the effect is due to the strong ability of IL-12 to stimulate the secretion of positive-acting IFN-γ by T cells or NK cells (16, 41). We did not try to block the effect of IL-12 in vivo by the administration of anti-IFN-γ because of the obvious effect that anti-IFN-γ would have had on elicitation of CS responses per se. However, and importantly, we observed almost comparable IL-12 reversal of in vitro proliferative responses of tolerized T cells in the presence of neutralizing Abs against either IFN-γ or IL-2 (Fig. 6). This strongly suggested a direct action of IL-12 on CS effector T cells in our established i.v. hapten tolerance system, as suggested by another system, in which IL-12 acted directly on Th1 clones to induce IL-2-independent proliferation (42). This direct effect of IL-12 on T cells probably synergizes with signals due to costimulatory molecules, without effects due to additional production of IFN-γ to induce T cell activation (36, 43). In fact, we observed that the reversing effects of IL-12 were blocked in vitro by anti-B7-1, more strongly than by anti-B7-2 mAb, and by both Abs (Fig. 7,A). Since in contrast to anti-B7-2 mAb, anti-B7-1 mAb had little effect on proliferative responses of nontolerized LNC (normally CS immunized) to haptenized APC (Fig. 7 B), there appeared to be some differences in responsiveness to these costimulatory molecule-mediated pathways between normal CS immune cells and tolerized CS effector T cells.
We do not know yet from our results how IL-12 synergizes with costimulatory molecule mediated signals to T cells, and how these pathways are different between CS immune cells and tolerized cells. IL-12 has been reported to be produced by monocytes and macrophages, and these cells function as important APC during normal immune responses. B7 molecules on these APC interact with CD28 and CTLA-4 on activated T cells, and these interactions are implicated in the priming of naive T cells, but not in the activation of memory T cells (43). However, some systems suggest that costimulation via B7 is required for functional activation of memory effector T cells (44), and IL-12 is also an important cytokine for activating memory T cells with costimulation via these molecules (43). These results are well correlated with our findings that costimulatory molecule-mediated pathways are required not only for full activation of CS effector cells (Fig. 7,B), but also, more importantly, for reactivation of tolerized cells by IL-12 (Fig. 7 A). As important roles for expression of B7-1 molecules and IL-12 in breaking peripheral tolerance have been implicated in other systems, such as in the pathogenesis of autoimmune diseases (45, 46) and in the augmentation of tumor immunity (47), the same mechanisms may be operative in our system to reverse tolerance by IL-12.
We also observed almost comparable reversal of impaired CS in established tolerance in vivo by administrating IL-12 systemically to mice at the time of elicitation of CS ear-swelling responses (Fig. 8). This effect was mimicked partially by in vivo administration of IFN-γ, although the magnitude of reversal of tolerance for CS ear-swelling responses was significantly greater for IL-12 (Fig. 8 vs Fig. 9). Moreover, we could not obtain further enhancement of CS ear-swelling responses by increasing the dose of IFN-γ (Fig. 9). These results again correlated well with our in vitro data, showing that IL-12 has an additional mechanism, beyond induction of IFN-γ, to reverse established CS tolerance by synergizing with B7-1 and B7-2 costimulatory molecule signal pathways.
In summary, these results taken together suggest that IL-12 probably acts by two possible pathways to produce a reversal of the established tolerance of CS that was induced previously by high i.v. doses of hapten. One pathway may have been the reversal of tolerance due to the production of a small amount of endogenous IFN-γ that is essential for elicitation of CS responses (4, 31, 32, 33). Our data indicate that it is unlikely that this small increase in IFN-γ occurred by IL-12 inhibiting counterbalancing Th2 responses. The other and probably major pathway of IL-12 reversal of established CS tolerance is proposed to have been due to a direct action of IL-12 on CS effector T cells to strengthen their responses, probably by increasing T cell costimulation mediated by signaling mechanisms acting via B7-1 and B7-2 (36, 43).
Finally, it has been suggested that IL-12 might be used clinically to augment impaired immune T cell function in clinical immunodeficiencies such as in HIV patients (48) or to augment tumor immunity (21, 47), in which reversing Ag-specific tolerance probably plays an important role. Thus, these current results may have important implications for the use of IL-12 to augment a broad range of impaired clinical immune responses, some of which may involve established tolerance.
The authors are indebted to Marilyn Avallone for her great secretarial skills, to laboratory colleagues for their advise and encouragement (Drs. W. Ptak, B. Nowak, G. Geba, R. Ramabhadran, V. Paliwal, and H. Sperl), and to Drs. S. E. Wolf, R. Coffman, and W. Waegell for their kind gifts of invaluable reagents.
This work was supported in part by grants from the National Institutes of Health to P.W.A. (AI12211, AI26689, and AI07174).
Abbreviations used in this paper: CS, contact sensitivity; PCl, picryl chloride (trinitrophenyl chloride); TNP, trintrophenyl; OX, oxazolone; TNBSA, trinitro-benzene sulfonic acid; LNC, lymph node cell.